zompist bboard

Under what circumstances do languages tend to borrow numerals? I'm asking about a few different situations here:

a) A language borrows a numeral for a number that lacked one in the native lexicon, e.g. "million" and related words in English
b) A language borrows a numeral for a number that already had one, and keeps both the original numeral and the foreign numeral for use in different contexts, e.g. most numerals in Japanese
c) A language borrows a numeral for a number that already had one, and it fully replaces the old numeral (I'm not aware of examples for this one, but wouldn't be surprised if it occurred somewhere)

Are there situations that generally cause a language to adopt new numerals? I recognize that many languages of hunter-gatherer cultures have historically had small sets of numerals prior to significant contact with colonizers or other representatives of higher-tech societies. Is there any pattern more specific than that?

Do you suppose numerals are different from other words that are borrowed? I can't think of a specific situation in which numerals would be borrowed, but other words would not. Numerals have their specific domain, but this domain includes other words (e.g. trade, or architecture, or whatever).


JAL

I'm wondering the same as the OP. What makes numerals different from other words is that they're function words, or at least I think they're more close to function than content words. One doesn't usually replace function words, do one?

English 3rd person pronouns have borrowings from Norse. Other function words do get borrowed on occasion if sometimes restricted in context (contra, per, are just two in English).

You can explore a lot of this for yourself over at my numbers page:

http://www.zompist.com/numbers.shtml

See also the sub-page on "Language information".

In brief: numbers are technology, and can be borrowed like any other technical term. You borrow them from the culture that has them.

Numbers are sometimes borrowed in a particular domain. An English example is the series ace, deuce, tray, cater, cinque, sice, first used for dice and later for cards, and borrowed from Old French. (Including "ace", from as.) A Quechua example is that the hours, and nothing else, are numbered in Spanish.

zompist wrote: A Quechua example is that the hours, and nothing else, are numbered in Spanish.

Doesn't inuktitut do the same thing?

Vietnamese has a set of Sinitic numerals and a set of native numerals. Strangely, only the ordinal numbers 1st and 4th are borrowed, but all the others are native. The Sinitic numerals are more common in compounds too, I think.

Numeral borrowing appears to have been fairly common in Californian languages, which is probably related to the fact that loanwords in the region were typically terms related to trading and trade goods.

Khmer borrowed the numbers 30, 40, 50, 60, 70, 80, 90 from Cantonese probably via Thai. I can't imagine Khmer lacked a word for 30, so it would in that case have to be an example of b). Larger numbers like 100, 1000, come from Thai. Unfortunately, I have no idea why this is.

zompist wrote:A Quechua example is that the hours, and nothing else, are numbered in Spanish.

Similar to Tagalog. I don't know exactly all the contexts in which the Spanish-derived numbers are used, but they are definitely used with time-telling. In general counting, the native Tagalog words are used

*edit*: Due to the very high amount of code-switching among Tagalog speakers, English numbers aren't infrequently used either.

If you look at specific language varieties, Standard English appears to have borrowed the pronunciation of 'one' and 'two' from English dialects. I've not managed to figure out which though. The sound changes involved seem similar to those in Borders Scots but with different intermediary sound changes so I would expect the sound changes are shared with parts of northern England, probably Northumbrian though I can't find much detail about Northumbrian online.

jmcd wrote:If you look at specific language varieties, Standard English appears to have borrowed the pronunciation of 'one' and 'two' from English dialects. I've not managed to figure out which though.

Online Etymology Dictionary wrote: Originally pronounced as it still is in only, and in dialectal good 'un, young 'un, etc.; the now-standard pronunciation "wun" began c.14c. in southwest and west England (Tyndale, a Gloucester man, spells it won in his Bible translation), and it began to be general 18c.

I haven't heard before that "two" would also be a borrowed pronunication. Can't find anything on-line about it.


JAL

zompist wrote:Numbers are sometimes borrowed in a particular domain. An English example is the series ace, deuce, tray, cater, cinque, sice, first used for dice and later for cards, and borrowed from Old French. (Including "ace", from as.) A Quechua example is that the hours, and nothing else, are numbered in Spanish.

In dialectal English, you also have the example of sheep-counting rhymes which use a series of numerals derived from Brythonic.

Has the origin of 'one' even been figured out yet?

(I worked that into Pembrish: OE ā raised and diphthongized before nasals, ejecting a semivowel when stressed and word-initial: aːn > oə̯n > woə̯n. Final nasals were later lost in most environments, and the centering diphthongs ea oa smoothed, so 'a' and 'one' are oa and woa, pronounced [ə] and [wɔ] -- Pembrish has linking n, so [ən wɔn] before vowels -- and 'only' would be woanlig [wɔnlitʃ] if i-mutation hadn't been preserved -- but it was, so it's eenlig [eːnlitʃ]. æ didn't break before nasals like a did, but e ē did, so the plural of monn is meann, pronounced [min].)

Nortaneous wrote:Has the origin of 'one' even been figured out yet?

You mean more figured out than the link I posted above?


JAL

jal wrote:

jmcd wrote:If you look at specific language varieties, Standard English appears to have borrowed the pronunciation of 'one' and 'two' from English dialects. I've not managed to figure out which though.

Online Etymology Dictionary wrote: Originally pronounced as it still is in only, and in dialectal good 'un, young 'un, etc.; the now-standard pronunciation "wun" began c.14c. in southwest and west England (Tyndale, a Gloucester man, spells it won in his Bible translation), and it began to be general 18c.

I haven't heard before that "two" would also be a borrowed pronunication. Can't find anything on-line about it.


JAL

Thanks, I should've thought to look there. 'two' does seem an unusual pronounciation; you would expect /two/ and a dialectal sound change of wo>u:. In any case, it's a similar sound change to that which occurred in Swedish.

zompist wrote:In brief: numbers are technology,

Aah, thanks!

jmcd wrote:Thanks, I should've thought to look there. 'two' does seem an unusual pronounciation; you would expect /two/ and a dialectal sound change of wo>u:. In any case, it's a similar sound change to that which occurred in Swedish.

"Two" isn't a borrowing - just a somewhat unusual sound change. Or rather, a combination of two changes. From the OED:

The pronunc[iation] (tuː), like that of who (huː) from OE. hwá, is due to labialization of the vowel by the w (cf. womb), which then disappeared before the related sound. The successive stages would thus be (twaː, twɔː, twoː, twuː, tuː).

In some northern Swedish dialects the English numeral two is also pronounce two, written most often as to. Compare Swedish två.
Funny, in standard Swedish there's a kid's rhyme or used to be one that has preserved a memory of this where två is pronounced as tu. It goes: ett, tu, tre (one, two, three).

Shrdlu wrote:In some northern Swedish dialects the English numeral two is also pronounce two, written most often as to. Compare Swedish två.
Funny, in standard Swedish there's a kid's rhyme or used to be one that has preserved a memory of this where två is pronounced as tu. It goes: ett, tu, tre (one, two, three).

Tu is an archaic form of två. Cf. tudelad, itu. Though tu alone can also be heard in some contexts.

Boşkoventi wrote:

jmcd wrote:Thanks, I should've thought to look there. 'two' does seem an unusual pronounciation; you would expect /two/ and a dialectal sound change of wo>u:. In any case, it's a similar sound change to that which occurred in Swedish.

"Two" isn't a borrowing - just a somewhat unusual sound change. Or rather, a combination of two changes. From the OED:

The pronunc[iation] (tuː), like that of who (huː) from OE. hwá, is due to labialization of the vowel by the w (cf. womb), which then disappeared before the related sound. The successive stages would thus be (twaː, twɔː, twoː, twuː, tuː).

Thanks, seeing the example of 'who' shows how that would be right.

A related question: If a language needs numerals, but has no language to borrow them from, how does/could it derive them?

Most often, from finger counting. (Or other body parts, but it's usually fingers.)

Small numbers may come from deictics ("this one", "that one", "the other one").

And of course, once you have a handful (ha) of numerals, you can start deriving new numerals from the old.

Doesn't have to be by agglutination either; you can also dualize or pluralize etc. things and then use those as new numerals.

zompist wrote: An English example is the series ace, deuce, tray, cater, cinque, sice, first used for dice and later for cards, and borrowed from Old French. (Including "ace", from as.)

Uzbek has that, too - when playing cards, Tajik / Farsi numbers (yek, du, she...) are used for counting, while otherwise Uzbek uses the inherited Turkic numbers (bir, ikki, u'ch...).

Qwynegold wrote:

Shrdlu wrote:In some northern Swedish dialects the English numeral two is also pronounce two, written most often as to. Compare Swedish två.
Funny, in standard Swedish there's a kid's rhyme or used to be one that has preserved a memory of this where två is pronounced as tu. It goes: ett, tu, tre (one, two, three).

Tu is an archaic form of två. Cf. tudelad, itu. Though tu alone can also be heard in some contexts.

Tu used to be the neuter form (and like neuter forms of adjectives, it had adverbial uses), so counting "ett, tu, tre" makes sense. Två comes from the Old Swedish masculine accusative form. The Old Swedish masculine nominative form was tvēr and later tvē.

The tve- in tvetydig and tvesala is an old prefixal form related to the greek and latin di- and bi- prefixes. Tvesala (selling an object to two different people) is regulated in 1 kap 5 § handelsbalken where an old dative form tvem can also be found: "Säljer man tvem ett; gälde skadan åter, (och böte tio daler,) och den behålle godset, som först köpte.".